In [68]:
data_plot (loaded_data,
box_Gaussian=True,
Plot_violinplot=True,
Dist_Plot=True,
Pair_plot=True,
Plot_confusionMatrix=True,
)
<Figure size 640x480 with 0 Axes>
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
!pip install seaborn
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/ Requirement already satisfied: seaborn in /usr/local/lib/python3.10/dist-packages (0.12.2) Requirement already satisfied: pandas>=0.25 in /usr/local/lib/python3.10/dist-packages (from seaborn) (1.5.3) Requirement already satisfied: numpy!=1.24.0,>=1.17 in /usr/local/lib/python3.10/dist-packages (from seaborn) (1.24.3) Requirement already satisfied: matplotlib!=3.6.1,>=3.1 in /usr/local/lib/python3.10/dist-packages (from seaborn) (3.7.1) Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (3.0.9) Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (4.39.3) Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (23.1) Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (1.4.4) Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (8.4.0) Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (2.8.2) Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (0.11.0) Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib!=3.6.1,>=3.1->seaborn) (1.0.7) Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=0.25->seaborn) (2022.7.1) Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib!=3.6.1,>=3.1->seaborn) (1.16.0)
pip install -U scikit-learn
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/ Requirement already satisfied: scikit-learn in /usr/local/lib/python3.10/dist-packages (1.2.2) Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn) (3.1.0) Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn) (1.2.0) Requirement already satisfied: scipy>=1.3.2 in /usr/local/lib/python3.10/dist-packages (from scikit-learn) (1.10.1) Requirement already satisfied: numpy>=1.17.3 in /usr/local/lib/python3.10/dist-packages (from scikit-learn) (1.24.3)
pip install scipy
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/ Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (1.10.1) Requirement already satisfied: numpy<1.27.0,>=1.19.5 in /usr/local/lib/python3.10/dist-packages (from scipy) (1.24.3)
import seaborn as sns
# sns.set(style="whitegrid")
import warnings
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.neural_network import MLPClassifier
from sklearn.svm import SVC, NuSVC
warnings.filterwarnings('ignore')
from sklearn.model_selection import train_test_split
# from tflearn.data_utils import to_categorical
from sklearn import preprocessing
from sklearn.model_selection import GridSearchCV, RandomizedSearchCV
from sklearn.metrics import confusion_matrix
from sklearn.decomposition import PCA
from sklearn.decomposition import FastICA
from keras.utils import to_categorical
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import SVC
from scipy import stats
from scipy.stats import uniform, randint
from sklearn.metrics import f1_score
from sklearn.model_selection import KFold, StratifiedKFold, RepeatedStratifiedKFold
from sklearn.metrics import roc_curve, auc, accuracy_score
from sklearn.metrics import classification_report
from scipy import interp
loaded_data = pd.read_csv('https://docs.google.com/spreadsheets/d/e/2PACX-1vTv03yjUMskcQXdIqAs800W-5T4EU8J31WQXwkQ9SQHU56f3mqZAxk5yx0d6PlLny4-KfF9QszeyHGu/pub?output=csv')
list_all_models=[] #every classifier auc values are stored in it
New_features= [] #Feature that are renamed
random_initializer=100 #random initializer
n_dots=50
##########################################################
for i in range(8):
#for renaming dataset of columns features F1 -- F8
New_features.append('F'+str(i+1))
##-------------------------------------------------##
# Gaussian distribution
def Box_Gaussian(loaded_data):
fig=plt.figure(figsize=(15,6)) #define figure size
fig.suptitle("Box Gaussian plot of all features")
n_scaler = preprocessing.StandardScaler() #standardization function
temp_Data = n_scaler.fit_transform(loaded_data) #pass into function for standrd.
for i in range(8): #loop for all 8 feature
plt.subplot(2, 4, i+1) #subplot for 2 rows in 4 columns
Data = temp_Data[:,i] #data for every feature
sns.kdeplot(Data, shade=True,color='red', alpha=0.3) #kernel density function under red shaded arae
ax = sns.boxplot(Data, saturation=0.9, color="green") #This gives green shaded area
plt.gca().invert_yaxis() #Y-Axis in reversed PyPlot
# plt.title('F'+str(i+1))
frame1 = plt.gca()
frame1.axes.xaxis.set_ticklabels([]) #removing xlabel data
plt.ylim((-0.5,0.65)) #y axis limit
plt.tight_layout() #This module provides routines to adjust subplot params so that subplots are nicely fit in the figure.
for patch in ax.artists:
r, g, b, a = patch.get_facecolor() #Get the facecolor of the Axes.
patch.set_facecolor((r, g, b, 0.3)) #set colour intensity
##-------------------------------------------------###
# Pairs plots are just showing all variables paired with all the other variables
def pair_plot(loaded_data):
plt.figure()
pair_plot =sns.pairplot(data=loaded_data,
height=3,
hue='Outcome',
diag_kind='kde')
pair_plot.fig.suptitle("Pairplot of all features")
plt.show()
##-------------------------------------------------#########
def plot_confusionMatrix(loaded_data):
'''
Parameters :
Input - data is the pandas type variable
Output -visualization of correalation matrix of
input data
'''
sns.set(font_scale=1.15) # Set aesthetic parameters in one step.
ax = plt.figure(figsize=(10, 8)) #set figure size https://seaborn.pyloaded_data.org/generated/seaborn.set.html
plt.title("Confusion Matrix of all features")
sns.heatmap(loaded_data.corr(), # input correlation matrix of dataset
vmax=1.0, #Values to anchor the colormap, otherwise they are inferred from
#the data and other keyword arguments.
vmin=0.0,
linewidths=0.01,
square=False, #If True, set the Axes aspect to “equal” so each cell will be square-shaped.
annot=True, #If True, write the data value in each cell.
linecolor="black") #Color of the lines that will divide each cell.
#https://seaborn.pyloaded_data.org/generated/seaborn.heatmap.html
b, t = plt.ylim() # discover the values for bottom and top
b += 0.5 # Add 0.5 to the bottom
t -= 0.5 # Subtract 0.5 from the top
plt.ylim(b, t) # update the ylim(bottom, top) values
plt.show()
##-------------------------------------------------##
# this function plot violin plot of
# every feature
def plot_violinplot (loaded_data):
'''
Parameters :
Input - data is the pandas type variable
Output - The violinplot plot for
eight feature of input data
'''
# A violin plot is a method of plotting numeric loaded_data.
# It is similar to box plot with a rotated kernel
# density plot on each side. Violin plots are similar
# to box plots, except that they also show the probability
# density of the data at different values (in the simplest
# case this could be a histogram).
fig, ax = plt.subplots(2,4, figsize=(12,6))
# #set numbers of rows and columns of subplot and figure size
sns.set()
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F1, ax=ax[0,0]) #violine plot for F1 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F2, ax=ax[0,1]) #violine plot for F2 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F3, ax=ax[0,2]) #violine plot for F3 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F4, ax=ax[0,3]) #violine plot for F4 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F5, ax=ax[1,0]) #violine plot for F5 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F6, ax=ax[1,1]) #violine plot for F6 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F7, ax=ax[1,2]) #violine plot for F7 feature
sns.violinplot(x = loaded_data.Outcome, y=loaded_data.F8, ax=ax[1,3]) #violine plot for F8 feature
fig.suptitle("Violin plot of all features")
fig.tight_layout()
# https://seaborn.pyloaded_data.org/generated/seaborn.violinplot.html
##-------------------------------------------------##
# Plots univariate distribution
def dist_Plot(loaded_data):
fig, ax = plt.subplots(2,4, figsize=(12,5)) #set numbers of rows and columns of subplot
sns.set()
sns.distplot(loaded_data.F1, bins = 10, ax=ax[0,0]) #Flexibly plot a univariate distribution of observations.
sns.distplot(loaded_data.F2, bins = 10, ax=ax[0,1])
sns.distplot(loaded_data.F3, bins = 10, ax=ax[0,2])
sns.distplot(loaded_data.F4, bins = 10, ax=ax[0,3])
sns.distplot(loaded_data.F5, bins = 10, ax=ax[1,0])
sns.distplot(loaded_data.F6, bins = 10, ax=ax[1,1])
sns.distplot(loaded_data.F7, bins = 10, ax=ax[1,2])
sns.distplot(loaded_data.F8, bins = 10, ax=ax[1,3])
fig.suptitle("Gaussian Distribution of all features")
fig.tight_layout() #This module provides routines to adjust subplot params
# so that subplots are nicely fit in the figure.
##-------------------------------------------------##
#Function for manually remove outleirs
def Manual (loaded_data):
max_Pregnancies = loaded_data.F1.max() #maximum feature of F1
data = loaded_data[loaded_data.F1!=max_Pregnancies] #find where extreme value is absent and remove extreme
max_Glucose = loaded_data.F2.max() #maximum feature of F2
loaded_data = loaded_data[loaded_data.F2!=max_Glucose] #find where extreme value is absent and remove
for i in range(4): #in this loop we succesively remove 4 minimum element
min_Glucose = loaded_data.F2.min() #find minimum
loaded_data = loaded_data[loaded_data.F2!=min_Glucose] #reject minimum
max_BloodPressure = loaded_data.F3.max() #maximum feature of F3
loaded_data = loaded_data[loaded_data.F3!=max_BloodPressure] #find where extreme value is absent and remove
for i in range(2): #in this loop we succesively remove 2 extreme element
max_skinthickness = loaded_data.F4.max()
loaded_data = loaded_data[loaded_data.F4!=max_skinthickness]
for i in range(25): #in this loop we succesively remove 25 extreme element
max_Insulin = loaded_data.F5.max()
loaded_data = loaded_data[loaded_data.F5!=max_Insulin]
max_bmi = loaded_data.F6.max()
loaded_data = loaded_data[loaded_data.F6!=max_bmi]
for i in range(4): #in this loop we succesively remove 4 minimum element
min_bmi = loaded_data.F6.min()
loaded_data = loaded_data[loaded_data.F6!=min_bmi]
for i in range(20): #in this loop we succesively remove 20 extreme element
max_DiabetesPedigreeF = loaded_data.F7.max()
loaded_data = loaded_data[loaded_data.F7!=max_DiabetesPedigreeF]
for i in range(20): #in this loop we succesively remove 20 extreme element
max_age = loaded_data.F8.max()
loaded_data = loaded_data[loaded_data.F8!=max_age]
df =loaded_data
return df
##-------------------------------------------------##
# Automatic outlair rejector
def IQR_Mean (loaded_data):
for i in range(8):
x = loaded_data[New_features[i]]
Q1 = x.quantile(0.25) # Q1 is the "middle" value in the first half of the rank-ordered data set.
Q3 = x.quantile(0.75) # Q3 is the "middle" value in the second half of the rank-ordered data set.
IQR = Q3-Q1 # The interquartile range is equal to Q3 minus Q1.
mean = x.mean() #mean of feature
for j in range(569): # loop for first 569 elements of feature
temp = x[j] # every feature value
LW = (Q1 - 1.5 * IQR) #lower considerable range of gaussian distribution
UW = (Q3 + 1.5 * IQR) #upper considerable range of gaussian distribution
if temp < LW: #replace upper value with mean
x[j] = mean
if temp > UW: #replace lower value with mean
x[j] = mean
loaded_data[New_features[i]] = x
return loaded_data
##-------------------------------------------------##
# Outlier rejector w.r.t median value
def IQR_Median (loaded_data):
for i in range(8):
x = loaded_data[New_features[i]]
Q1 = x.quantile(0.25)
Q3 = x.quantile(0.75)
IQR = Q3-Q1
median = x.quantile(0.5) # find the median
for j in range(569): #replace the first 569 values with respect to median
temp = x[j]
LW = (Q1 - 1.5 * IQR)
UW = (Q3 + 1.5 * IQR)
if temp < LW: #replace upper value with median
x[j] = median
if temp > UW:
x[j] = median #replace upper value with median
loaded_data[New_features[i]] = x
return loaded_data
##-------------------------------------------------##
# Outlier rejector with 1.5 times * IQR; significant in gaussian distribution
def IQR (loaded_data):
for i in range(8): # for every feature
Q1 = loaded_data[New_features[i]].quantile(0.25)
Q3 = loaded_data[New_features[i]].quantile(0.75)
IQR = Q3-Q1 #find IQR
LW = (Q1 - 1.5 * IQR) #find lower boundary
# print(LW)
UW = (Q3 + 1.5 * IQR) #find upper boundary
# print(UW)
loaded_data = loaded_data[loaded_data[New_features[i]]<UW] #drop greater than upper limit
loaded_data = loaded_data[loaded_data[New_features[i]]>LW] #drop smaller than lower limit
return data
##-------------------------------------------------##
#Condition based outlier rejector function
def outlier_Rejection (data, iqr_Mean, iqr_Medain, iqr, manual):
if iqr_Mean == True: #reject outleir with Mean
data = IQR_Mean (loaded_data)
if iqr_Medain == True: #reject outleir with Median
data = IQR_Medain (loaded_data)
if iqr == True: #reject outleir in IQR range
data = IQR (loaded_data)
if manual == True: #reject outleir with manual
data = Manual (loaded_data)
return data
##-------------------------------------------------##
#Plots based on I/P condition
def data_plot (data,
Pair_plot,
Dist_Plot,
Plot_violinplot,
Plot_confusionMatrix,
box_Gaussian):
if Pair_plot ==True:
pair_plot(loaded_data)
if Dist_Plot ==True:
dist_Plot(loaded_data)
if Plot_violinplot ==True:
plot_violinplot (loaded_data)
if Plot_confusionMatrix ==True:
plot_confusionMatrix(loaded_data)
if box_Gaussian ==True:
Box_Gaussian(loaded_data)
##-------------------------------------------------##
def replace_zero(temp_Data, field, target):
mean_by_target = loaded_data.loc[loaded_data[field] != 0, [field, target]].groupby(target).mean()
loaded_data.loc[(loaded_data[field] == 0)&(loaded_data[target] == 0), field] = mean_by_target.iloc[0][0]
loaded_data.loc[(loaded_data[field] == 0)&(loaded_data[target] == 1), field] = mean_by_target.iloc[1][0]
##-------------------------------------------------##
def metrics (y_true, y_pred, probas_):
points=n_dots*'-'
print(points)
fpr, tpr, thresholds = roc_curve(y_true, probas_[:, 1])
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
roc_auc = auc(fpr, tpr)
print("Detailed classification report for current fold:")
print()
print(classification_report(y_true, y_pred))
print()
print("Area Under ROC (AUC): {}".format(roc_auc))
print()
print('Confusion Matrix for current fold: ')
print(confusion_matrix(y_true, y_pred))
print()
print("Accuracy for Current Fold: {}".format(accuracy_score(y_true, y_pred)))
print()
tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
return tn, fp, fn, tp, roc_auc, fpr, tpr
##-------------------------------------------------##
def average_ROC(mean_fpr,tprs,aucs,TP,TN,FP,FN):
sen = (np.sum(TP))/(np.sum(TP)+np.sum(FN))
spe = (np.sum(TN))/(np.sum(TN)+np.sum(FP))
mean_tpr = np.mean(tprs, axis=0)
mean_tpr[-1] = 1.0
mean_auc = np.mean(aucs)
std_auc = np.std(aucs)
ax = plt.axes()
ax.grid(color='lightgray', linestyle='-', linewidth=.5)
ax.set_facecolor("white")
ax.spines['bottom'].set_color('#000000')
ax.spines['top'].set_color('#000000')
ax.spines['right'].set_color('#000000')
ax.spines['left'].set_color('#000000')
plt.plot(mean_fpr, mean_tpr, color='blue',
label=r'Avg. ROC (AUC (avg $\pm$ std) = %0.3f $\pm$ %0.3f)' % (mean_auc, std_auc),
lw=2, alpha=.8)
plt.scatter((1-spe), sen, s=80, c='r', marker='x',)
plt.scatter(0, sen, s=80, c='r', marker='x',)
plt.scatter((1-spe),0, s=80, c='r', marker='x',)
plt.axhline(y=sen, color='r', linestyle='--')
plt.axvline(x=(1-spe), color='r', linestyle='--')
plt.text((1-spe), 0.02, "FPR={:0.3f}".format((1-spe)))
plt.text(0.009, sen+0.05, "TPR={:0.3f}".format(sen))
std_tpr = np.std(tprs, axis=0)
tprs_upper = np.minimum(mean_tpr + std_tpr, 1)
tprs_lower = np.maximum(mean_tpr - std_tpr, 0)
plt.fill_between(mean_fpr, tprs_lower, tprs_upper, color='darkgray', alpha=0.5,
label=r'$\pm$ 1 Standard deviation')
plt.xticks(np.arange(0.0, 1.01, step=0.1))
plt.yticks(np.arange(0.0, 1.01, step=0.1))
left=0.0
right=1.0
plt.xlim(left, right)
plt.ylim(left, right)
plt.xlabel('False Positive Rate (FPR)')
plt.ylabel('True Positive Rate (TPR)')
plt.legend(loc="lower right")
plt.show()
##-------------------------------------------------##
def plot_Current_ROC(fpr,tpr,iterator,roc_auc):
plt.plot(fpr,
tpr,
alpha=0.35,
linewidth=1)
##-------------------------------------------------##
def New_Model (classifier, X_Train, Y_Train, tuned_parameters, verbose):
clf = GridSearchCV(classifier,
tuned_parameters,
verbose=verbose,
cv=5,
scoring='roc_auc',
n_jobs=-1)
clf.fit(X_Train, Y_Train)
return clf
##-------------------------------------------------##
def average_performance(aucs,Accuracy,TP,TN,FP,FN):
print()
n_dotsav=(n_dots-len('Average'))//2
print('-'*n_dotsav+'Average'+'-'*n_dotsav)
print("AUC (Avg. +/- Std.) is %0.3f +/- %0.3f" %(np.mean(aucs),np.std(aucs)))
print("Accuracy (Avg. +/- Std.) is %0.3f +/- %0.3f" %(np.mean(Accuracy),np.std(Accuracy)))
cm = [[int(np.mean(TP)), int(np.mean(FP))],[int(np.mean(FN)), int(np.mean(TN))]]
print ('Avg. CM is '+str(cm))
cm = [[int(np.sum(TP)), int(np.sum(FP))],[int(np.sum(FN)), int(np.sum(TN))]]
print ('Total for all folds CM is '+str(cm))
re_auc=str(round(np.mean(aucs), 3))+'+/-'+str(round(np.std(aucs),3))
list_all_models.append(re_auc)
##-------------------------------------------------##
#Feature selection based on algorithms
def feature_Selector(data, algo, n_feature):
if algo =='corr':
if n_feature ==4:
data = loaded_data[['F2','F5','F4','F6','Outcome']]
return loaded_data.iloc[:,:4].values, loaded_data.iloc[:,4:].values
if n_feature ==6:
data = loaded_data[['F1','F2','F4','F5','F6','F8','Outcome']]
return loaded_data.iloc[:,:6].values, loaded_data.iloc[:,6:].values
if algo == 'ICA':
X_Data= loaded_data.iloc[:,:8].values
ICA = FastICA(n_components=n_feature, random_state=12)
X_Data = ICA.fit_transform(X_Data)
return X_Data , loaded_data.iloc[:,8:].values
if algo=='PCA':
X_Data= loaded_data.iloc[:,:8].values
pca = PCA(n_components=n_feature)
X_Data = pca.fit_transform(X_Data)
return X_Data , loaded_data.iloc[:,8:].values
if algo == 'None':
return loaded_data.iloc[:,:8].values, loaded_data.iloc[:,8:].values
loaded_data.head(n=6)
| Pregnancies | Glucose | BloodPressure | SkinThickness | Insulin | BMI | DiabetesPedigreeFunction | Age | Outcome | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | 6 | 148 | 72 | 35 | 0 | 33.6 | 0.627 | 50 | 1 |
| 1 | 1 | 85 | 66 | 29 | 0 | 26.6 | 0.351 | 31 | 0 |
| 2 | 8 | 183 | 64 | 0 | 0 | 23.3 | 0.672 | 32 | 1 |
| 3 | 1 | 89 | 66 | 23 | 94 | 28.1 | 0.167 | 21 | 0 |
| 4 | 0 | 137 | 40 | 35 | 168 | 43.1 | 2.288 | 33 | 1 |
| 5 | 5 | 116 | 74 | 0 | 0 | 25.6 | 0.201 | 30 | 0 |
| New Name | Original Name | Comments | |
|---|---|---|---|
| F1 | Pregnancies | Number of times pregnant | |
| F2 | Glucose | Plasma glucose concentration a 2 hours in an oral glucose tolerance test | |
| F3 | BloodPressure | Diastolic blood pressure (mm Hg) | |
| F4 | SkinThickness | Triceps skin fold thickness (mm) | |
| F5 | Insulin | 2-Hour serum insulin (mu U/ml) | |
| F6 | BMI | Body mass index (weight in kg/(height in m)^2) | |
| F7 | DiabetesPedigreeFunction | Diabetes pedigree function | |
| F8 | Age | Age (years) |
loaded_data = pd.DataFrame({'F1':loaded_data.iloc[:,:8].values[:,0],
'F2':loaded_data.iloc[:,:8].values[:,1],
'F3':loaded_data.iloc[:,:8].values[:,2],
'F4':loaded_data.iloc[:,:8].values[:,3],
'F5':loaded_data.iloc[:,:8].values[:,4],
'F6':loaded_data.iloc[:,:8].values[:,5],
'F7':loaded_data.iloc[:,:8].values[:,6],
'F8':loaded_data.iloc[:,:8].values[:,7],
'Outcome':loaded_data.iloc[:,8:].values[:,0]})
# Now shows different values in Outcome
loaded_data['Outcome'].value_counts()
0 500 1 268 Name: Outcome, dtype: int64
loaded_data.shape
(768, 9)
loaded_data.describe()
| F1 | F2 | F3 | F4 | F5 | F6 | F7 | F8 | Outcome | |
|---|---|---|---|---|---|---|---|---|---|
| count | 768.000000 | 768.000000 | 768.000000 | 768.000000 | 768.000000 | 768.000000 | 768.000000 | 768.000000 | 768.000000 |
| mean | 3.845052 | 120.894531 | 69.105469 | 20.536458 | 79.799479 | 31.992578 | 0.471876 | 33.240885 | 0.348958 |
| std | 3.369578 | 31.972618 | 19.355807 | 15.952218 | 115.244002 | 7.884160 | 0.331329 | 11.760232 | 0.476951 |
| min | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.078000 | 21.000000 | 0.000000 |
| 25% | 1.000000 | 99.000000 | 62.000000 | 0.000000 | 0.000000 | 27.300000 | 0.243750 | 24.000000 | 0.000000 |
| 50% | 3.000000 | 117.000000 | 72.000000 | 23.000000 | 30.500000 | 32.000000 | 0.372500 | 29.000000 | 0.000000 |
| 75% | 6.000000 | 140.250000 | 80.000000 | 32.000000 | 127.250000 | 36.600000 | 0.626250 | 41.000000 | 1.000000 |
| max | 17.000000 | 199.000000 | 122.000000 | 99.000000 | 846.000000 | 67.100000 | 2.420000 | 81.000000 | 1.000000 |
data_plot (loaded_data,
box_Gaussian=True,
Plot_violinplot=True,
Dist_Plot=True,
Pair_plot=True,
Plot_confusionMatrix=True,
)
<Figure size 640x480 with 0 Axes>
| Process Selector | Preprocessing | |
|---|---|---|
| P | Outlier Rejection | |
| Q | Filling Missing Value | |
| R | Standardization |
Each P, Q, and R process has four techniques for feature selection which are N/A, PCA, ICA, and correlation-based feature selection
print('Before removing outliers : ' + str(loaded_data.shape))
##-------------------------------------------------################
loaded_data = outlier_Rejection (loaded_data,
iqr_Mean=False,
iqr_Medain=False,
iqr=True,
manual=False)
print('After outlier were removed : ' + str(loaded_data.shape))
Before removing outliers : (768, 9) After outlier were removed : (768, 9)
loaded_data.isna().sum()
Pregnancies 0 Glucose 0 BloodPressure 0 SkinThickness 0 Insulin 0 BMI 0 DiabetesPedigreeFunction 0 Age 0 Outcome 0 dtype: int64
X_loaded_data,Target_Variable = feature_Selector(loaded_data, algo='None', n_feature=6)
print('Shape After Feature Selection: ' + str(loaded_data.shape))
Shape After Feature Selection: (768, 9)
Defining K-fold values; 10 here
kf = StratifiedKFold(n_splits=10,
shuffle=True,
random_state=random_initializer)
iterator=0
FP = [] # For False Positive
FN = [] # Stores False Negative
TP = [] # To store True Positive
TN = [] # To store True Negative
tprs = [] # True positive
Accuracy = [] # Stores Accuracy
aucs_kNN = [] # Stores AUC
mean_fpr = np.linspace(0, 1, 100)
fig = plt.figure(figsize=(12, 6))
for train_index, test_index in kf.split(X_loaded_data,Target_Variable):
X_Train, X_Test = X_loaded_data[train_index], X_loaded_data[test_index]
Y_Train, Y_Test = Target_Variable[train_index], Target_Variable[test_index]
# Hyperparameters
n_neighbors = [1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49]
leaf_size = [5,10,15,20,25,30,35,40,45,50]
Distance = [1,2]
tuned_parameters = [ {'n_neighbors': n_neighbors,
'algorithm' : ['brute'],
'p':Distance},
{'n_neighbors': n_neighbors,
'algorithm' : ['ball_tree'],
'leaf_size' : leaf_size,
'p':Distance},
{'n_neighbors': n_neighbors,
'algorithm' : ['kd_tree'],
'leaf_size' : leaf_size,
'p':Distance}]
clf = New_Model (classifier = KNeighborsClassifier(), #create the model
X_Train = X_Train,
Y_Train = Y_Train,
tuned_parameters = tuned_parameters,
verbose=0)
tn, fp, fn, tp, roc_auc, fpr, tpr = metrics (y_true = Y_Test, #get the values of model evaluation
y_pred = clf.predict(X_Test),
probas_ = clf.predict_proba(X_Test))
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
aucs_kNN.append(roc_auc)
plot_Current_ROC (fpr,tpr,iterator,roc_auc) #plot the roc of current fold
iterator += 1
TN.append(tn)
FP.append(fp)
FN.append(fn)
TP.append(tp)
Accuracy.append(accuracy_score(Y_Test, clf.predict(X_Test)))
average_ROC(mean_fpr,tprs,aucs_kNN,TP,TN,FP,FN) #plot average roc curve
average_performance(aucs_kNN,Accuracy,TP,TN,FP,FN) #print the average performance of the model
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.71 0.84 0.77 50
1 0.56 0.37 0.44 27
accuracy 0.68 77
macro avg 0.63 0.61 0.61 77
weighted avg 0.66 0.68 0.66 77
Area Under ROC (AUC): 0.7188888888888889
Confusion Matrix for current fold:
[[42 8]
[17 10]]
Accuracy for Current Fold: 0.6753246753246753
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.92 0.83 50
1 0.75 0.44 0.56 27
accuracy 0.75 77
macro avg 0.75 0.68 0.69 77
weighted avg 0.75 0.75 0.73 77
Area Under ROC (AUC): 0.8277777777777777
Confusion Matrix for current fold:
[[46 4]
[15 12]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.94 0.83 50
1 0.79 0.41 0.54 27
accuracy 0.75 77
macro avg 0.77 0.67 0.68 77
weighted avg 0.76 0.75 0.73 77
Area Under ROC (AUC): 0.8274074074074075
Confusion Matrix for current fold:
[[47 3]
[16 11]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.68 0.82 0.75 50
1 0.47 0.30 0.36 27
accuracy 0.64 77
macro avg 0.58 0.56 0.55 77
weighted avg 0.61 0.64 0.61 77
Area Under ROC (AUC): 0.7233333333333334
Confusion Matrix for current fold:
[[41 9]
[19 8]]
Accuracy for Current Fold: 0.6363636363636364
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.98 0.85 50
1 0.92 0.41 0.56 27
accuracy 0.78 77
macro avg 0.84 0.69 0.71 77
weighted avg 0.81 0.78 0.75 77
Area Under ROC (AUC): 0.8292592592592591
Confusion Matrix for current fold:
[[49 1]
[16 11]]
Accuracy for Current Fold: 0.7792207792207793
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.77 0.96 0.86 50
1 0.87 0.48 0.62 27
accuracy 0.79 77
macro avg 0.82 0.72 0.74 77
weighted avg 0.81 0.79 0.77 77
Area Under ROC (AUC): 0.8807407407407406
Confusion Matrix for current fold:
[[48 2]
[14 13]]
Accuracy for Current Fold: 0.7922077922077922
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.80 0.96 0.87 50
1 0.88 0.56 0.68 27
accuracy 0.82 77
macro avg 0.84 0.76 0.78 77
weighted avg 0.83 0.82 0.81 77
Area Under ROC (AUC): 0.8166666666666667
Confusion Matrix for current fold:
[[48 2]
[12 15]]
Accuracy for Current Fold: 0.8181818181818182
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.79 0.90 0.84 50
1 0.75 0.56 0.64 27
accuracy 0.78 77
macro avg 0.77 0.73 0.74 77
weighted avg 0.78 0.78 0.77 77
Area Under ROC (AUC): 0.8337037037037036
Confusion Matrix for current fold:
[[45 5]
[12 15]]
Accuracy for Current Fold: 0.7792207792207793
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.72 0.86 0.78 50
1 0.56 0.35 0.43 26
accuracy 0.68 76
macro avg 0.64 0.60 0.61 76
weighted avg 0.66 0.68 0.66 76
Area Under ROC (AUC): 0.8046153846153846
Confusion Matrix for current fold:
[[43 7]
[17 9]]
Accuracy for Current Fold: 0.6842105263157895
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.84 0.84 0.84 50
1 0.69 0.69 0.69 26
accuracy 0.79 76
macro avg 0.77 0.77 0.77 76
weighted avg 0.79 0.79 0.79 76
Area Under ROC (AUC): 0.8034615384615386
Confusion Matrix for current fold:
[[42 8]
[ 8 18]]
Accuracy for Current Fold: 0.7894736842105263
---------------------Average--------------------- AUC (Avg. +/- Std.) is 0.807 +/- 0.047 Accuracy (Avg. +/- Std.) is 0.746 +/- 0.057 Avg. CM is [[12, 4], [14, 45]] Total for all folds CM is [[122, 49], [146, 451]]
iterator=0
FP = [] # For False Positive
FN = [] # Stores False Negative
TP = [] # To store True Positive
TN = [] # To store True Negative
tprs = [] # True positive
Accuracy = [] # Stores Accuracy
aucs_Tree = [] # Stores AUC
mean_fpr = np.linspace(0, 1, 100)
fig = plt.figure(figsize=(12, 6))
for train_index, test_index in kf.split(X_loaded_data,Target_Variable):
X_Train, X_Test = X_loaded_data[train_index], X_loaded_data[test_index]
Y_Train, Y_Test = Target_Variable[train_index], Target_Variable[test_index]
# Hyperparameters
tuned_parameters = {'criterion': ['gini','entropy'],
'splitter': ['best'],
'min_samples_split':[0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0],
'min_samples_leaf': [1,2,3,4,5] }
clf = New_Model (classifier = DecisionTreeClassifier( random_state=random_initializer),
X_Train = X_Train, # create the model using DecisionTree Classifier
Y_Train = Y_Train,
tuned_parameters = tuned_parameters,
verbose=0)
tn, fp, fn, tp, roc_auc, fpr, tpr = metrics (y_true = Y_Test, #get the evaluation parameters of model
y_pred = clf.predict(X_Test),
probas_ = clf.predict_proba(X_Test))
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
aucs_Tree.append(roc_auc) # plot the roc of current fold
iterator += 1
TN.append(tn)
FP.append(fp)
FN.append(fn)
TP.append(tp)
Accuracy.append(accuracy_score(Y_Test, clf.predict(X_Test)))
average_ROC(mean_fpr,tprs,aucs_Tree,TP,TN,FP,FN) #plot average roc curve
average_performance(aucs_Tree,Accuracy,TP,TN,FP,FN) #print the average performance of the model
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.78 0.86 0.82 50
1 0.68 0.56 0.61 27
accuracy 0.75 77
macro avg 0.73 0.71 0.72 77
weighted avg 0.75 0.75 0.75 77
Area Under ROC (AUC): 0.8129629629629629
Confusion Matrix for current fold:
[[43 7]
[12 15]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.78 0.84 0.81 50
1 0.65 0.56 0.60 27
accuracy 0.74 77
macro avg 0.71 0.70 0.70 77
weighted avg 0.73 0.74 0.73 77
Area Under ROC (AUC): 0.8281481481481482
Confusion Matrix for current fold:
[[42 8]
[12 15]]
Accuracy for Current Fold: 0.7402597402597403
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.88 0.72 0.79 50
1 0.61 0.81 0.70 27
accuracy 0.75 77
macro avg 0.74 0.77 0.74 77
weighted avg 0.78 0.75 0.76 77
Area Under ROC (AUC): 0.8329629629629629
Confusion Matrix for current fold:
[[36 14]
[ 5 22]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.76 0.78 0.77 50
1 0.58 0.56 0.57 27
accuracy 0.70 77
macro avg 0.67 0.67 0.67 77
weighted avg 0.70 0.70 0.70 77
Area Under ROC (AUC): 0.727037037037037
Confusion Matrix for current fold:
[[39 11]
[12 15]]
Accuracy for Current Fold: 0.7012987012987013
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.77 0.94 0.85 50
1 0.81 0.48 0.60 27
accuracy 0.78 77
macro avg 0.79 0.71 0.73 77
weighted avg 0.79 0.78 0.76 77
Area Under ROC (AUC): 0.851111111111111
Confusion Matrix for current fold:
[[47 3]
[14 13]]
Accuracy for Current Fold: 0.7792207792207793
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.87 0.82 0.85 50
1 0.70 0.78 0.74 27
accuracy 0.81 77
macro avg 0.79 0.80 0.79 77
weighted avg 0.81 0.81 0.81 77
Area Under ROC (AUC): 0.9048148148148147
Confusion Matrix for current fold:
[[41 9]
[ 6 21]]
Accuracy for Current Fold: 0.8051948051948052
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.80 0.86 0.83 50
1 0.70 0.59 0.64 27
accuracy 0.77 77
macro avg 0.75 0.73 0.73 77
weighted avg 0.76 0.77 0.76 77
Area Under ROC (AUC): 0.8488888888888889
Confusion Matrix for current fold:
[[43 7]
[11 16]]
Accuracy for Current Fold: 0.7662337662337663
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.84 0.64 0.73 50
1 0.54 0.78 0.64 27
accuracy 0.69 77
macro avg 0.69 0.71 0.68 77
weighted avg 0.74 0.69 0.70 77
Area Under ROC (AUC): 0.7514814814814814
Confusion Matrix for current fold:
[[32 18]
[ 6 21]]
Accuracy for Current Fold: 0.6883116883116883
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.82 0.74 0.78 50
1 0.58 0.69 0.63 26
accuracy 0.72 76
macro avg 0.70 0.72 0.71 76
weighted avg 0.74 0.72 0.73 76
Area Under ROC (AUC): 0.8188461538461538
Confusion Matrix for current fold:
[[37 13]
[ 8 18]]
Accuracy for Current Fold: 0.7236842105263158
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.87 0.68 0.76 50
1 0.57 0.81 0.67 26
accuracy 0.72 76
macro avg 0.72 0.74 0.72 76
weighted avg 0.77 0.72 0.73 76
Area Under ROC (AUC): 0.81
Confusion Matrix for current fold:
[[34 16]
[ 5 21]]
Accuracy for Current Fold: 0.7236842105263158
---------------------Average--------------------- AUC (Avg. +/- Std.) is 0.819 +/- 0.048 Accuracy (Avg. +/- Std.) is 0.743 +/- 0.034 Avg. CM is [[17, 10], [9, 39]] Total for all folds CM is [[177, 106], [91, 394]]
iterator=0
FP = [] # For False Positive
FN = [] # Stores False Negative
TP = [] # To store True Positive
TN = [] # To store True Negative
tprs = [] # True positive
Accuracy = [] # Stores Accuracy
aucs_Forest = [] # Stores AUC
mean_fpr = np.linspace(0, 1, 100)
fig = plt.figure(figsize=(12, 6))
for train_index, test_index in kf.split(X_loaded_data,Target_Variable):
X_Train, X_Test = X_loaded_data[train_index], X_loaded_data[test_index]
Y_Train, Y_Test = Target_Variable[train_index], Target_Variable[test_index]
# Hyperparameters
tuned_parameters = {'criterion': ['gini','entropy']}
clf = New_Model (classifier = RandomForestClassifier( random_state=random_initializer),
X_Train = X_Train, # create a model using random forest classifier
Y_Train = Y_Train,
tuned_parameters = tuned_parameters,
verbose=0)
tn, fp, fn, tp, roc_auc, fpr, tpr = metrics (y_true = Y_Test, #evaluation parameters of random forest model
y_pred = clf.predict(X_Test),
probas_ = clf.predict_proba(X_Test))
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
aucs_Forest.append(roc_auc)
plot_Current_ROC (fpr,tpr,iterator,roc_auc) #plot the roc of current fold
iterator += 1
TN.append(tn)
FP.append(fp)
FN.append(fn)
TP.append(tp)
Accuracy.append(accuracy_score(Y_Test, clf.predict(X_Test)))
average_ROC(mean_fpr,tprs,aucs_Forest,TP,TN,FP,FN) #plot average roc curve
average_performance(aucs_Forest,Accuracy,TP,TN,FP,FN) #print the average performance of the model
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.76 0.82 0.79 50
1 0.61 0.52 0.56 27
accuracy 0.71 77
macro avg 0.68 0.67 0.67 77
weighted avg 0.71 0.71 0.71 77
Area Under ROC (AUC): 0.764074074074074
Confusion Matrix for current fold:
[[41 9]
[13 14]]
Accuracy for Current Fold: 0.7142857142857143
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.76 0.88 0.81 50
1 0.68 0.48 0.57 27
accuracy 0.74 77
macro avg 0.72 0.68 0.69 77
weighted avg 0.73 0.74 0.73 77
Area Under ROC (AUC): 0.8059259259259259
Confusion Matrix for current fold:
[[44 6]
[14 13]]
Accuracy for Current Fold: 0.7402597402597403
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.77 0.86 0.81 50
1 0.67 0.52 0.58 27
accuracy 0.74 77
macro avg 0.72 0.69 0.70 77
weighted avg 0.73 0.74 0.73 77
Area Under ROC (AUC): 0.8333333333333333
Confusion Matrix for current fold:
[[43 7]
[13 14]]
Accuracy for Current Fold: 0.7402597402597403
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.73 0.80 0.76 50
1 0.55 0.44 0.49 27
accuracy 0.68 77
macro avg 0.64 0.62 0.63 77
weighted avg 0.66 0.68 0.67 77
Area Under ROC (AUC): 0.7185185185185184
Confusion Matrix for current fold:
[[40 10]
[15 12]]
Accuracy for Current Fold: 0.6753246753246753
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.92 0.83 50
1 0.75 0.44 0.56 27
accuracy 0.75 77
macro avg 0.75 0.68 0.69 77
weighted avg 0.75 0.75 0.73 77
Area Under ROC (AUC): 0.8433333333333333
Confusion Matrix for current fold:
[[46 4]
[15 12]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.85 0.90 0.87 50
1 0.79 0.70 0.75 27
accuracy 0.83 77
macro avg 0.82 0.80 0.81 77
weighted avg 0.83 0.83 0.83 77
Area Under ROC (AUC): 0.9099999999999999
Confusion Matrix for current fold:
[[45 5]
[ 8 19]]
Accuracy for Current Fold: 0.8311688311688312
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.81 0.88 0.85 50
1 0.74 0.63 0.68 27
accuracy 0.79 77
macro avg 0.78 0.75 0.76 77
weighted avg 0.79 0.79 0.79 77
Area Under ROC (AUC): 0.8770370370370371
Confusion Matrix for current fold:
[[44 6]
[10 17]]
Accuracy for Current Fold: 0.7922077922077922
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.81 0.88 0.85 50
1 0.74 0.63 0.68 27
accuracy 0.79 77
macro avg 0.78 0.75 0.76 77
weighted avg 0.79 0.79 0.79 77
Area Under ROC (AUC): 0.8385185185185184
Confusion Matrix for current fold:
[[44 6]
[10 17]]
Accuracy for Current Fold: 0.7922077922077922
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.79 0.82 0.80 50
1 0.62 0.58 0.60 26
accuracy 0.74 76
macro avg 0.71 0.70 0.70 76
weighted avg 0.73 0.74 0.73 76
Area Under ROC (AUC): 0.8415384615384616
Confusion Matrix for current fold:
[[41 9]
[11 15]]
Accuracy for Current Fold: 0.7368421052631579
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.88 0.84 0.86 50
1 0.71 0.77 0.74 26
accuracy 0.82 76
macro avg 0.79 0.80 0.80 76
weighted avg 0.82 0.82 0.82 76
Area Under ROC (AUC): 0.8276923076923077
Confusion Matrix for current fold:
[[42 8]
[ 6 20]]
Accuracy for Current Fold: 0.8157894736842105
---------------------Average--------------------- AUC (Avg. +/- Std.) is 0.826 +/- 0.051 Accuracy (Avg. +/- Std.) is 0.759 +/- 0.046 Avg. CM is [[15, 7], [11, 43]] Total for all folds CM is [[153, 70], [115, 430]]
SVM
from sklearn.svm import SVC
iterator=0
FP = [] # For False Positive
FN = [] # Stores False Negative
TP = [] # To store True Positive
TN = [] # To store True Negative
tprs = [] # True positive
Accuracy = [] # Stores Accuracy
aucs_svm = [] # Stores AUC
mean_fpr = np.linspace(0, 1, 100)
fig = plt.figure(figsize=(12, 6))
for train_index, test_index in kf.split(X_loaded_data,Target_Variable):
X_Train, X_Test = X_loaded_data[train_index], X_loaded_data[test_index]
Y_Train, Y_Test = Target_Variable[train_index], Target_Variable[test_index]
# Hyperparameters
tuned_parameters = [{'kernel': ['rbf','linear']}]
##-------------------------------------------------###
clf = New_Model (classifier = SVC(random_state=0,probability=True),
X_Train = X_Train,
Y_Train = Y_Train,
tuned_parameters=tuned_parameters,
verbose=0)
tn, fp, fn, tp, roc_auc, fpr, tpr = metrics (y_true = Y_Test, # model evaluation parameters
y_pred = clf.predict(X_Test),
probas_ = clf.predict_proba(X_Test))
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
aucs_svm.append(roc_auc)
plot_Current_ROC (fpr,tpr,iterator,roc_auc) #plot the roc of current fold
iterator += 1
TN.append(tn)
FP.append(fp)
FN.append(fn)
TP.append(tp)
Accuracy.append(accuracy_score(Y_Test, clf.predict(X_Test)))
average_ROC(mean_fpr,tprs,aucs_svm,TP,TN,FP,FN) #plot average roc curve
average_performance(aucs_svm,Accuracy,TP,TN,FP,FN) #print the average performance of the model
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.86 0.80 50
1 0.65 0.48 0.55 27
accuracy 0.73 77
macro avg 0.70 0.67 0.68 77
weighted avg 0.72 0.73 0.72 77
Area Under ROC (AUC): 0.7925925925925925
Confusion Matrix for current fold:
[[43 7]
[14 13]]
Accuracy for Current Fold: 0.7272727272727273
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.86 0.80 50
1 0.65 0.48 0.55 27
accuracy 0.73 77
macro avg 0.70 0.67 0.68 77
weighted avg 0.72 0.73 0.72 77
Area Under ROC (AUC): 0.808888888888889
Confusion Matrix for current fold:
[[43 7]
[14 13]]
Accuracy for Current Fold: 0.7272727272727273
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.78 0.92 0.84 50
1 0.78 0.52 0.62 27
accuracy 0.78 77
macro avg 0.78 0.72 0.73 77
weighted avg 0.78 0.78 0.77 77
Area Under ROC (AUC): 0.8762962962962964
Confusion Matrix for current fold:
[[46 4]
[13 14]]
Accuracy for Current Fold: 0.7792207792207793
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.73 0.76 0.75 50
1 0.52 0.48 0.50 27
accuracy 0.66 77
macro avg 0.63 0.62 0.62 77
weighted avg 0.66 0.66 0.66 77
Area Under ROC (AUC): 0.7318518518518519
Confusion Matrix for current fold:
[[38 12]
[14 13]]
Accuracy for Current Fold: 0.6623376623376623
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.78 0.94 0.85 50
1 0.82 0.52 0.64 27
accuracy 0.79 77
macro avg 0.80 0.73 0.75 77
weighted avg 0.80 0.79 0.78 77
Area Under ROC (AUC): 0.8607407407407408
Confusion Matrix for current fold:
[[47 3]
[13 14]]
Accuracy for Current Fold: 0.7922077922077922
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.84 0.96 0.90 50
1 0.90 0.67 0.77 27
accuracy 0.86 77
macro avg 0.87 0.81 0.83 77
weighted avg 0.86 0.86 0.85 77
Area Under ROC (AUC): 0.9118518518518518
Confusion Matrix for current fold:
[[48 2]
[ 9 18]]
Accuracy for Current Fold: 0.8571428571428571
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.83 0.86 0.84 50
1 0.72 0.67 0.69 27
accuracy 0.79 77
macro avg 0.77 0.76 0.77 77
weighted avg 0.79 0.79 0.79 77
Area Under ROC (AUC): 0.8807407407407407
Confusion Matrix for current fold:
[[43 7]
[ 9 18]]
Accuracy for Current Fold: 0.7922077922077922
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.80 0.88 0.84 50
1 0.73 0.59 0.65 27
accuracy 0.78 77
macro avg 0.76 0.74 0.75 77
weighted avg 0.77 0.78 0.77 77
Area Under ROC (AUC): 0.8348148148148148
Confusion Matrix for current fold:
[[44 6]
[11 16]]
Accuracy for Current Fold: 0.7792207792207793
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.77 0.82 0.80 50
1 0.61 0.54 0.57 26
accuracy 0.72 76
macro avg 0.69 0.68 0.68 76
weighted avg 0.72 0.72 0.72 76
Area Under ROC (AUC): 0.8184615384615385
Confusion Matrix for current fold:
[[41 9]
[12 14]]
Accuracy for Current Fold: 0.7236842105263158
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.85 0.88 0.86 50
1 0.75 0.69 0.72 26
accuracy 0.82 76
macro avg 0.80 0.79 0.79 76
weighted avg 0.81 0.82 0.81 76
Area Under ROC (AUC): 0.823076923076923
Confusion Matrix for current fold:
[[44 6]
[ 8 18]]
Accuracy for Current Fold: 0.8157894736842105
---------------------Average--------------------- AUC (Avg. +/- Std.) is 0.834 +/- 0.049 Accuracy (Avg. +/- Std.) is 0.766 +/- 0.053 Avg. CM is [[15, 6], [11, 43]] Total for all folds CM is [[151, 63], [117, 437]]
iterator=0
FP = [] # For False Positive
FN = [] # Stores False Negative
TP = [] # To store True Positive
TN = [] # To store True Negative
tprs = [] # True positive
Accuracy = [] # Stores Accuracy
aucs_NB = [] # Stores AUC
mean_fpr = np.linspace(0, 1, 100)
fig = plt.figure(figsize=(12, 6))
for train_index, test_index in kf.split(X_loaded_data,Target_Variable):
X_Train, X_Test = X_loaded_data[train_index], X_loaded_data[test_index]
Y_Train, Y_Test = Target_Variable[train_index], Target_Variable[test_index]
# Hyperparameters
var_smoothing = [1e-01,
1e-02,
1e-03,
1e-04,
1e-05,
1e-06,
1e-07,
1e-08,
1e-09,
1e-10,
1e-11,
1e-12]
tuned_parameters = [{'var_smoothing': var_smoothing}]
##-------------------------------------------------###
clf = New_Model (classifier = GaussianNB(), # create model using Naive Bias
X_Train = X_Train,
Y_Train = Y_Train,
tuned_parameters = tuned_parameters,
verbose=0)
tn, fp, fn, tp, roc_auc, fpr, tpr = metrics (y_true = Y_Test, # model evaluation parameters
y_pred = clf.predict(X_Test),
probas_ = clf.predict_proba(X_Test))
tprs.append(interp(mean_fpr, fpr, tpr))
tprs[-1][0] = 0.0
aucs_NB.append(roc_auc)
plot_Current_ROC (fpr,tpr,iterator,roc_auc) #plot the roc of current fold
iterator += 1
TN.append(tn)
FP.append(fp)
FN.append(fn)
TP.append(tp)
Accuracy.append(accuracy_score(Y_Test, clf.predict(X_Test)))
average_ROC(mean_fpr,tprs,aucs_NB,TP,TN,FP,FN) #plot average roc curve
average_performance(aucs_NB,Accuracy,TP,TN,FP,FN) #print the average performance of the model
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.73 0.82 0.77 50
1 0.57 0.44 0.50 27
accuracy 0.69 77
macro avg 0.65 0.63 0.64 77
weighted avg 0.68 0.69 0.68 77
Area Under ROC (AUC): 0.7622222222222221
Confusion Matrix for current fold:
[[41 9]
[15 12]]
Accuracy for Current Fold: 0.6883116883116883
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.82 0.82 0.82 50
1 0.67 0.67 0.67 27
accuracy 0.77 77
macro avg 0.74 0.74 0.74 77
weighted avg 0.77 0.77 0.77 77
Area Under ROC (AUC): 0.8503703703703704
Confusion Matrix for current fold:
[[41 9]
[ 9 18]]
Accuracy for Current Fold: 0.7662337662337663
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.80 0.86 0.83 50
1 0.70 0.59 0.64 27
accuracy 0.77 77
macro avg 0.75 0.73 0.73 77
weighted avg 0.76 0.77 0.76 77
Area Under ROC (AUC): 0.817037037037037
Confusion Matrix for current fold:
[[43 7]
[11 16]]
Accuracy for Current Fold: 0.7662337662337663
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.74 0.68 0.71 50
1 0.48 0.56 0.52 27
accuracy 0.64 77
macro avg 0.61 0.62 0.61 77
weighted avg 0.65 0.64 0.64 77
Area Under ROC (AUC): 0.6881481481481482
Confusion Matrix for current fold:
[[34 16]
[12 15]]
Accuracy for Current Fold: 0.6363636363636364
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.75 0.92 0.83 50
1 0.75 0.44 0.56 27
accuracy 0.75 77
macro avg 0.75 0.68 0.69 77
weighted avg 0.75 0.75 0.73 77
Area Under ROC (AUC): 0.8451851851851852
Confusion Matrix for current fold:
[[46 4]
[15 12]]
Accuracy for Current Fold: 0.7532467532467533
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.81 0.92 0.86 50
1 0.80 0.59 0.68 27
accuracy 0.81 77
macro avg 0.80 0.76 0.77 77
weighted avg 0.80 0.81 0.80 77
Area Under ROC (AUC): 0.8733333333333333
Confusion Matrix for current fold:
[[46 4]
[11 16]]
Accuracy for Current Fold: 0.8051948051948052
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.84 0.86 0.85 50
1 0.73 0.70 0.72 27
accuracy 0.81 77
macro avg 0.79 0.78 0.78 77
weighted avg 0.80 0.81 0.80 77
Area Under ROC (AUC): 0.8711111111111112
Confusion Matrix for current fold:
[[43 7]
[ 8 19]]
Accuracy for Current Fold: 0.8051948051948052
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.86 0.84 0.85 50
1 0.71 0.74 0.73 27
accuracy 0.81 77
macro avg 0.79 0.79 0.79 77
weighted avg 0.81 0.81 0.81 77
Area Under ROC (AUC): 0.7992592592592592
Confusion Matrix for current fold:
[[42 8]
[ 7 20]]
Accuracy for Current Fold: 0.8051948051948052
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.74 0.78 0.76 50
1 0.52 0.46 0.49 26
accuracy 0.67 76
macro avg 0.63 0.62 0.62 76
weighted avg 0.66 0.67 0.67 76
Area Under ROC (AUC): 0.7976923076923077
Confusion Matrix for current fold:
[[39 11]
[14 12]]
Accuracy for Current Fold: 0.6710526315789473
--------------------------------------------------
Detailed classification report for current fold:
precision recall f1-score support
0 0.87 0.90 0.88 50
1 0.79 0.73 0.76 26
accuracy 0.84 76
macro avg 0.83 0.82 0.82 76
weighted avg 0.84 0.84 0.84 76
Area Under ROC (AUC): 0.8338461538461539
Confusion Matrix for current fold:
[[45 5]
[ 7 19]]
Accuracy for Current Fold: 0.8421052631578947
---------------------Average--------------------- AUC (Avg. +/- Std.) is 0.814 +/- 0.053 Accuracy (Avg. +/- Std.) is 0.754 +/- 0.064 Avg. CM is [[15, 8], [10, 42]] Total for all folds CM is [[159, 80], [109, 420]]
print('Finished')
Finished
%%shell
jupyter nbconvert --to html Final.ipynb